Dennis Geller wrote:
Thanks. Very helpful. However, having cleared that (and lots more
that derived from it) up, I think I'm good for now (with one more
question, below)
First, in case it helps anyone else, after following your suggestions
I found that art of my problem/confusion was that I had managed to
not have parallel directory hierarchies under descriptors and src --
that made attempting to replicate the tutorial example unlikely to
succeed. When I straightened that out I was unable to run until I
added the directory that contains the TypeSystem.xml file to the class
path. I don't believe the example needed that, but I sure did.
UIMA Descriptors have an "<import>" capability, to include other
descriptors within them. Many of the examples use this, for instance,
to include a common type system specification. There are 2 ways to
refer to what you are importing - by "location" or by "name". If you do
things by "name", then UIMA uses Java's ability to locate resources by
name in the Classpath (or UIMA Datapath). If your descriptors import
the type system "by name", then it is best to include it in your class path.
I just almost had a successful run. However, it coughed because a file
had a "non-XML character, 0x0."
Where was this character? If it was in your XML descriptors, then that
needs to be corrected. It is possible to analyze arbitrary data,
including "byte" data containing any characters, in UIMA; see
http://incubator.apache.org/uima/downloads/releaseDocs/2.2.1-incubating/docs/html/tutorials_and_users_guides/tutorials_and_users_guides.html#ugr.tug.aas.sofa_data_formats
This also happened when i was running the unmodified tutorial example.
Can you say where this character occurred in the unmodified tutorial example
-Marshall
